Mapping from frames to delexicalized sentences is generally done by encoder-decoder models. The input to the encoder is a sequence of tokens Xt that represent the dialogue act and its arguments. The encoder reads all the input slot/value representations, and the decoder outputs delexicalized sentence.

Encoder-Decoder Model

There are models that help the mapping from frames to delexicalized sentences.

Hamilton College

Delexicalization is the process of replacing specific words in the training set that represent slot values with a generic placeholder token representing the slot, as shown below:

Delexicalization

An on-going but a helpful book resource about NLP
https://web.stanford.edu/~jurafsky/slp3/

Speech and Language Processing (3rd ed. draft) 

Transition Models

The opposite process of delexicalization.

Learn Before

Related

Learn After